MisAnthropic Statement Says Anthropic Is Not a Safe AI Company Ahead of IPO

MisAnthropic Statement Says Anthropic Is Not a Safe AI Company Ahead of IPO

A new open letter called the MisAnthropic Statement is taking direct aim at Anthropic, the company behind Claude, just as it prepares for its initial public offering. Published on November 18, 2026, and backed by the Center for AI Safety (CAIS), the statement argues that Anthropic uses safety branding to win over investors and the public while racing to build self-improving AI. Its signatories include Steven Pinker, Judea Pearl, Hugging Face co-founder Thomas Wolf and George Hotz.

What the MisAnthropic Statement Says

The core text is short, just three sentences. It says Anthropic warns that recursively self-improving AI could concentrate power and escape human control, yet races to build it first. It ends with a blunt line: “Anthropic is not a safe AI company.”

Recursive self-improvement means AI systems getting better at improving themselves, including helping design the next generation of models. Many safety researchers see it as one of the riskiest paths in AI, because humans could lose track of what the systems are doing.

The Three Main Complaints

CAIS spelled out the coalition’s reasoning in its announcement. Here is how the accusations break down:

AccusationWhat the coalition claims
Racing toward self-improving AIAnthropic warns about the risks while pitching automated AI research to investors as a strength
Safety as marketingThe “AI safety and research company” label helps recruit talent and court investors
Weaker commitmentsAnthropic loosened its Responsible Scaling Policy in February and lobbied against several state AI bills

These are the coalition’s claims, not established findings. The CAIS release says Anthropic dropped its unilateral pause commitments and lobbied to weaken bills such as SB 1047 and the RAISE Act, while privately backing federal pre-emption of state AI laws. The CAIS announcement does not include a response from Anthropic.

Who Signed the MisAnthropic Statement

What makes this letter unusual is how much its signers normally disagree. The coalition mixes AI safety researchers, open-source advocates, AI ethicists and even accelerationists. Notable names on the statement’s website include:

  • Steven Pinker, Harvard psychologist
  • Judea Pearl, Turing Award winner at UCLA
  • Oliver Hart, Nobel laureate in economics
  • Laurence Tribe and Lawrence Lessig of Harvard Law School
  • Thomas Wolf, co-founder of Hugging Face
  • Igor Babuschkin, xAI co-founder
  • Max Tegmark of MIT and the Future of Life Institute
  • Guillaume Verdon, known online as “Beff Jezos,” founder of the e/acc movement
  • George Hotz, founder of comma.ai and the tiny corp
  • Fields Medalists Peter Scholze, James Maynard and David Mumford

CAIS executive director Dan Hendrycks said the group disagrees on almost everything but shares concern about a company that believes its own dominance is the greater good.

Why the Anthropic IPO Timing Matters

The letter lands as Anthropic heads toward public markets, a step we have been tracking since its revenue and IPO plans surfaced. An IPO puts a company’s story in front of millions of new investors. By attacking the safety reputation that sets Anthropic apart from rivals, the coalition is trying to shape how that story is told.

It is not the only pressure Anthropic faces right now. Earlier this month, a group of nonprofits sued major labs, including Anthropic, to freeze frontier AI work.

What This Means for You

If you use Claude every day, nothing changes about the product today. The letter is a public pressure campaign, not a regulatory action or a lawsuit. Still, it is worth paying attention to for a few reasons.

  • Safety promises from AI companies are becoming a key part of how they sell themselves, so expect more scrutiny of them.
  • If you are thinking about investing in AI stocks, read the safety and risk sections of filings, not just the headlines.
  • Debates over self-improving AI are moving from research circles into mainstream politics and policy.

For more background on the company, see our Claude overview and our wider AI news coverage.

Takeaway: a rare alliance of critics from opposite sides of the AI debate is trying to strip Anthropic of its “safe lab” image right before its biggest financial moment.

Sources

Primary source: Center for AI Safety announcement. Full statement and signatory list at the MisAnthropic Statement website.

Conclusion

The MisAnthropic Statement is one of the sharpest public challenges Anthropic has faced, and its timing ahead of the IPO is clearly deliberate. By uniting safety advocates, open-source leaders and accelerationists, CAIS is arguing that criticism of Anthropic crosses the usual battle lines. Whether it changes anything depends on how investors, regulators and Anthropic itself respond. For now, it puts the company’s safety reputation under a bright spotlight at exactly the moment that reputation matters most.

Newsletter

New AI tools in your inbox

The best new tools and the AI news that matters, in one short email. Unsubscribe any time.

Free · No spam · Unsubscribe any time

✓ Looks like you’re subscribed

Loading Google’s sign-up form…

Looks like you’re already subscribed

Thanks for reading! New AI tools and the AI news that matters will land in your inbox.

Not subscribed yet? Refresh the page to load the sign-up form.

The sign-up form couldn’t load. If you use an ad or content blocker, allow news.google.com and reload the page.

Free instant check before you pay. Takes a few seconds.

Scroll to Top